There’s also one more thing that would need to be addressed:
But such polygons already are drawn on a map, e.g. on Wikipedia… And I believe the process of making the outline drawn on the map on Wikipedia is automatic (due to the Wiki-Article and OSM-Element being linked by a Wikidata identifier?). But there’s currently no established way to tell any data consumer “but please don’t draw THIS geometry in particular, on account that it’s an intentionally vague representation”
Isn’t the issue that the entire class of feature is fuzzy? Fortunately, Wikipedia gives editors a good deal of control over which OSM feature to highlight and how. A Wikipedian selecting (or allowing) the Alps polygon to highlight as a well-defined feature is simply making a mistake on their end. Even then, the infobox maps default to a zoom level that fits the whole feature, so in the case of the Alps, it’s such a low zoom level that there doesn’t seem to be much of a problem?
“The Santa Cruz Mountains” (south of San Francisco, California) are a single way, defined by quirky, hyper-local “peaks” (160 of them) and that’s being specific with what we mean by these. OK.
Yes, I am one among its several authors, though for the low-hundreds-of-kilometers long-thing that this is, that’s a terse, even smart, AND accurate way to map it in OSM. Decades later, it works; version 60, going strong, lookin’ good.
If you want to avoid OSM, there is also
Mountain ranges would belong into this file, pull requests welcome:
Way: Great Dividing Range (211234843) | OpenStreetMap (discord discussion)
The linked discussion talked about increasing the geometry detail, either based on an area high elevation, another based a manually drawn polygon over the slope map, another based on the water catchments / drainage maps.
In general though I think we should be open to including these rough geometry features, whether they are mountain ranges, mountains, named places. Tags to indicate the geometry is subjective or in-precises are a good idea.
While we might not all agree on the exact boundaries, I’ve found that through community discussions we usually end up with a description of the bounds accepted by most and usually better than what any one individual would come up with, and that’s the whole point of OSM, collaboration can produce something collectively superior.
I can help writing a wiki page listing current practices (from country to country) + currently used tags (place=region + Key:region:type - OpenStreetMap Wiki is also used 1700 times not listed on this thread) + project making use of it, and when are polygons or way or relation or nodes preferred.
imho, one cause for this variety, is unrelated to OSM, it is how different human cultures consider those vague “named mountain things”
I’ll explicitly take France (where I’m from) in example : we have an history of naming “regions” that are not related to administrative entities, and mountains (specifically the French Alps where I live) also have vagued named areas we call “massifs” (but I’m still over simplifying, from place to place, we also of “chaîne” “chainon” “crêtes” which relates to the shape).
What I want to express here, is that having as main goal “world uniformity“ might be met with opposition.
This is only speculation on my part, but instead of accepting those “mountain regions” to have varying geometries, vague boundaries and inconsistencies, people (in OSM but also wikipedia or more officially from SOIUSA - Wikipedia) have tried too much to define modeling rules for the sake of universality.
At least for France (because don’t worry, this also is a mess !), my belief is that naming a “massif” is a culture, not a science. In other words : we should not have a computer, geology or physical approach to the question : “from where to where does a massif/range/crest extends”
Because we risk to needlessly force into :
My current choice is to decide on geometries type and shape from local and historical usage and try to approximate the shape without over-noding it by first choosing the statistical dominant answer to : “Is X in or out of massif/range/valley Y” and building the shape accordingly.
For cross referencing, I started in 2017 a wiki page (sorry in french and only about French Alps, but raising those type of questions and a draft for “current practices” for osm tagging) : FR:France/Régions Naturelles et Massifs montagneux - OpenStreetMap Wiki